Jay Fallon
Founding Engineer
Platform Architect
Full Stack Software Engineer
Tool Builder
Principal-level engineer who designs and operates multi-service production platforms end-to-end. Currently building Knokr, a music-industry platform managing 60,000+ artists, 1,500+ festivals, 1,000+ venues, and 25,000+ events across a federated architecture: three Next.js apps, a Python/FastAPI prediction service, and four BullMQ background workers sharing a single Postgres/Redis/S3 backbone.
View ResumeAbout
I'm a principal-level engineer who designs and operates multi-service production platforms end-to-end. Twenty-plus years of engineering experience across enterprise (Dassault Systèmes, Bose, State Street, Cengage) and consumer health (Form Health), with interdisciplinary education — a BS in Business Administration (International Business, Finance) from Northeastern and an MLA in Digital Media Design from Harvard Extension School.
Currently building Knokr, a music-industry platform managing 60,000+ artists, 1,500+ festivals, 1,000+ venues, and 25,000+ events. Federated architecture from day one: three Next.js apps (admin CMS, public site, media library) and a Python/FastAPI prediction service, all backed by four BullMQ workers sharing a single Postgres/Redis/S3 backbone. Solo architect and engineer — sole owner of schema, services, infrastructure, and the ~2,800 tests that hold it all together.
Deep on backend architecture, data modeling, and applied ML. pgvector with HNSW indexes power sub-100ms semantic search, CLIP image search, and "more like this" recommendations across the platform. A graph engine over 3.3M festival co-occurrence edges runs Node2Vec embeddings and Louvain community detection to surface emerging artists. Database-driven routing (URLs defined in the database, not the filesystem), six-tier RBAC with Redis-cached permissions delivering 90–95% latency reduction, and a multi-strategy data ingestion pipeline aggregating festival, venue, and event data from dozens of sources.
A long front-end track record informs how I build APIs and design systems — but the work I'm doing now is engineering at the foundation: schema design, service boundaries, queue topology, vector indexes, graph algorithms, and operational ergonomics. I'm seeking founding engineer, principal IC, or staff backend roles where one engineer can own architecture and ship across the stack.
Experience
Sole architect and engineer of a federated music-industry platform spanning a consumer-facing application and an operator/CMS backend. Chose multi-service over monolith from day one: separate Next.js apps, a Python/FastAPI prediction service, and BullMQ workers sharing Postgres, Redis, and S3. Designed the schema for a complex domain — entity resolution across artists, venues, festivals, and events; slug stability; ownership transfer; multi-tenant isolation; six-tier RBAC with three-layer route protection. pgvector + HNSW for sub-100ms semantic search. Node2Vec + Louvain on 3.3M co-occurrence edges for artist discovery. Anthropic Claude API for AI festival recommendations and lineup extraction from posters. ~2,800 tests across the system (Vitest, Playwright, Testcontainers, Supertest, Memlab).
- PostgreSQL
- Prisma
- pgvector
- Redis
- BullMQ
- Node.js
- Python
- FastAPI
- Next.js 16
- React 19
- TypeScript
- Tailwind
- Clerk
- Stripe
- AWS S3
- CloudFront
- Sentry
- Railway
- Anthropic Claude
- OpenAI
Rewrote the front-end from scratch for the site redesign and built a componentized design-system library used across the platform. Worked closely with marketing and product on the customer-facing platform for a medical weight-loss company; built and maintained the marketing site infrastructure and shared component library. Implemented Storybook documentation and supporting tooling to standardize component usage and reduce front-end fragmentation across teams.
- Next.js
- TypeScript
- Tailwind CSS
- Prismic
- Storybook
Senior IC role despite the manager title — hands-on engineer working alongside director-level peers on the SOLIDWORKS Cloud Platform. Implemented and maintained UI features across products to enforce design-system consistency and reduce accumulated technical debt. Designed and built a new content management system serving the entire product suite, replacing fragmented per-product docs. Partnered with new product teams on consumer-facing UIs while keeping product and design language consistent across the portfolio.
- Angular.js
- JavaScript
- CSS
- Cassandra
Contractor across two products in the SOLIDWORKS Apps for Kids suite. Rewrote the entire front-end in Angular 1.x; built a componentized design system to unify the separate applications. Consolidated disparate API streams into one structure for easier maintenance. Rewrote the end-to-end test suite with Mocha, Chai, and Puppeteer. Sole engineer on Classroom — a teacher-facing app for 30,000 students aged 8–14 with COPPA-compliant data and account flows. Containerized an Express suite of 12 interdependent services using Docker Compose, replacing manual per-service startup.
- Angular.js
- JavaScript
- Node.js
- Express
- Docker
- Docker Compose
- Mocha
- Chai
- Puppeteer
- Cassandra
Engineering and consulting roles spanning financial services, education, consumer audio, and pharmaceuticals. Full-stack web development, design system architecture, and front-end engineering at scale. Stack depth from this period includes six years of Angular.js, four years of Ruby on Rails, multi-year Express.js applications, and the full pre-React CSS lineage (CSS, SCSS, PostCSS, Stylus). Clients and employers via direct hire and consulting: Bose Corporation, State Street Global Advisors, Cengage Learning, Boston Technologies, netNumina Solutions (Fidelity, State Street, Johnson & Johnson, Pfizer, ADQSR), and Style Me Pretty.
- Angular.js
- Ruby on Rails
- Express.js
- JavaScript
- Node.js
- CSS
- SCSS
- PostCSS
- Stylus
Education
Completed Master of Liberal Arts degree in Digital Media Design through Harvard Extension School, focusing on user experience design, interaction design, and digital media production. Developed comprehensive understanding of design principles, user research methodologies, and digital product development.
- UX Design
- Digital Media
- Interaction Design
- User Research
Completed Bachelor of Science degree in Business Administration with majors in International Business and Finance, focusing on global operations and marketing. Participated in Northeastern's renowned cooperative education program, completing assignments at Boston's Better Snacks, FMRC, and PictureTel, gaining hands-on experience in operations, marketing, technical documentation, and internationalization.
- International Business
- Finance
- Operations
- Marketing
Projects
Active
Knokr
The public-facing knokr.com — a music-industry discovery and social platform consolidating 60,000+ artists, 1,500+ festivals, 1,000+ venues, and 25,000+ events into one searchable surface. City-context homepage threads through every section — Hottest Ticket, What's On, Hottest Venues, Latest Updated, Social Events — so picking a city reshapes the whole site. Discovery via semantic vector search and full-text PostgreSQL indexes; CLIP-based image search across user-contributed albums. Calendar with RSVPs ties into a BullMQ-backed reminder pipeline that DMs attendees at T−7d, T−1d, and T−1h in the venue's local timezone. Albums link explicitly to entities (artists, festivals, venues, cities, events) with NSFW + near-duplicate moderation, three-tier contribution policies, and save-to-album that re-uses S3 keys. User-to-user messaging gated by mutual-friend checks, with image upload through CloudFront. Animated Leaflet festival map, dream lineup builder, geographic radius-based discovery, and a conversational recommendation engine round out the consumer surface.
- Next.js 16
- React 19
- TypeScript
- PostgreSQL
- pgvector
- Prisma
- Redis
- Claude AI API
- @dnd-kit
- html-to-image
- Railway
Knokr Base
The administrative backbone of the Knokr ecosystem — a multi-tenant back office managing the full lifecycle of data across 60,000+ artists, 1,500+ festivals, 1,000+ venues, and 25,000+ events. Each user gets their own site at /f/[username]; entities (venues, artists, festivals, sponsors) manage their own branded presence with database-driven routing where URLs are defined in the database rather than file paths. Six-tier RBAC (SUPER_ADMIN, ADMIN, PUBLISHER, EDITOR, AUTHOR, READER) with three-layer route protection (middleware → layout → component) and Redis-cached permissions delivering 90–95% latency reductions. Four BullMQ workers handle graph computation, CLIP and OpenAI embeddings, insights aggregation, and social-card generation as independent Railway services. Includes Road Trips for multi-city tour planning, Fan Clubs auto-created per entity, Commemorative Tickets for verified attendance, an Entity Request workflow, and a Page Builder with draft/publish. Shares its Postgres/Redis/S3 infrastructure with knokr-lineups (public site) and knokr-media (photo platform).
- Next.js 16
- React 19
- TypeScript
- PostgreSQL
- pgvector
- Prisma
- Redis
- BullMQ
- OpenAI API
- Clerk
- AWS S3
- Railway
Knokr Experiments
Standalone applications built to test emerging technologies and validate ideas before integrating them into the Knokr platform. Each experiment is scoped to a specific problem, built fast, and deployed independently.
- Next.js
- TypeScript
- Redis
- CSS
- Canvas
In Development
Knokr Media
Integrated into knokr.com after testing
A photo and video album platform for the Knokr ecosystem — every album links explicitly to a Knokr entity (artist, festival, venue, or city), turning scattered fan media into structured visual coverage of the music industry. Standalone Next.js app sharing production Postgres, Redis, Clerk, and S3 with knokr-base and knokr-lineups so it deploys and iterates in isolation on Railway. Every uploaded photo runs through a CLIP embedding pipeline served by a separate worker service, powering semantic text search, "more like this" recommendations, NSFW moderation, and near-duplicate detection against pgvector HNSW indexes. Includes JIT Clerk profile sync, three-tier album contribution policies, denormalized rating rollups, golden-ratio client-side cropping, and a save-to-album action that re-uses S3 keys instead of copying objects. Covered by 376 tests across web and worker suites.
- Next.js 16
- React 19
- TypeScript
- PostgreSQL
- pgvector
- Prisma 6
- Redis
- BullMQ
- HeroUI
- Tailwind CSS 4
- Clerk
- AWS S3
- @xenova/transformers
- Railway
Mojo Boston
Demo — awaiting final content and visual design
A turnkey festival website exploring PayloadCMS as a white-label foundation for festival organizers. Built for a client's one-day festival, the project started from Payload's stock website template (blog + 4 basic blocks) and required substantial rework — 24 custom block components, a complete S3 media system replacing Payload's broken upload pipeline, live database integration pulling lineup, schedule, and venue data from Knokr's shared PostgreSQL infrastructure, interactive Leaflet venue maps, and a brand-specific design system. The experiment validated both the value of the Knokr data layer for powering external sites and the limitations of adapting a generic CMS for domain-specific use cases — informing the architecture for purpose-built festival site tooling.
- Next.js 16
- PayloadCMS 3.80
- TypeScript
- PostgreSQL
- AWS S3
- Leaflet
- Tailwind CSS 4
- Lexical
- Sharp
- Railway
Knokr Predictor
Ongoing development
A festival lineup prediction system that generates probable artist bookings for upcoming festival editions using Knokr's music discovery graph. Queries 3.3M pre-computed artist co-occurrence edges weighted across eight signal types (shared festivals, shared venues, band members, local/national scene overlap), combined with Louvain community detection scoring and genre-depth multipliers to rank candidates against a festival's existing lineup profile. Features automatic similar festival discovery, weighted random sampling with square-root flattening for variety, confidence tiering, and per-artist contributing factor breakdowns. Python prediction engine (FastAPI, asyncpg) communicating with a Next.js frontend via Redis queue, both reading from the shared Knokr PostgreSQL database. Deployed to Railway.
- FastAPI
- asyncpg
- Redis
- Next.js 16
- React 19
- TypeScript
- Prisma
- PostgreSQL
- HeroUI
- Tailwind CSS 4
- Railway
Knokr Fetch
Ongoing development
An intelligent event data extraction system that automatically harvests, verifies, deduplicates, and stages concert and tour event data from artist websites across the music industry. Python FastAPI backend with a multi-strategy extraction pipeline: JSON-LD structured data first, then platform widget detection (Bandsintown, Songkick, Seated), custom HTML parsing, and Claude LLM fallback for unstructured pages. Headless Chrome via Selenium for JavaScript-rendered content. Parallel job processing with asyncio semaphore-gated concurrency, WebSocket real-time progress streaming, and subprocess isolation per artist. Processes 52K artists with 99%+ accuracy on structured sources, handles 13+ international date formats, and deduplicates by content composite key rather than URL. PostgreSQL for production data, SQLite for local job persistence.
- Python
- FastAPI
- asyncio
- Selenium
- BeautifulSoup 4
- Anthropic Claude API
- httpx
- psycopg2
- SQLite
- WebSocket
- PostgreSQL
Inactive
Orchestra
Live — no longer under active development
A focused experiment in mapping artist relationships within the music industry. Enables users to search across 50,000+ artists and contribute band member connections through a curated submission process, building primary, secondary, and tertiary relationship networks. Built as a standalone platform to rapidly test community-driven data contribution workflows and graph-based artist discovery — learnings that directly informed the comparison models and artist relationship features in Knokr Base.
- Next.js 16
- React 19
- TypeScript
- PostgreSQL
- Prisma
- HeroUI
- Tailwind CSS
- Clerk
- Cloudflare Turnstile
- Railway
Nuncio
Live — no longer under active development
An open event curation platform built to test AI-powered data extraction and public content distribution at speed. Users create and share unlimited event lists with AI-assisted tour date extraction from images, embeddable widgets, and export formats including JSON-LD, iCal, and RSS feeds. Served as a proving ground for the image extraction pipeline and structured data patterns later adopted in Knokr's lineup extraction system.
- Next.js 16
- TypeScript
- PostgreSQL
- Prisma
- Redis
- BullMQ
- HeroUI
- Tailwind CSS
- Claude AI API
- Clerk
- AWS S3
Connections Plus
No longer under active development
A daily word puzzle game inspired by NYT Connections, featuring a multi-level progression system where red herring words accumulate across four difficulty tiers. Full-stack build with an admin interface for AI-assisted puzzle creation, calendar-based content scheduling, and anonymous player progress tracking. A fun side project that still exercises the full stack — auth, state management, content tooling, and polished UI with animations.
- Next.js 16
- TypeScript
- Redis
- HeroUI
- Tailwind CSS
- Claude AI API
- Framer Motion
- Vercel
Knokr — Earlier Iterations
No longer under active development
The original Knokr platform: a frontend for social music discovery, a back office supporting festival scheduling from single-day showcases to multi-day, multi-city, multi-stage events with 100+ artists, and KnokrFetch — a Python-based scraping system that pulled event data from Bandsintown, Songkick, and Seated alongside custom parsers for Wix, WordPress, Squarespace, and bespoke sites. These earlier builds established the domain model, data acquisition patterns, and scheduling complexity that the current Knokr and Knokr Base architecture is built on.
- Next.js
- Python
- FastAPI
- Selenium
- BeautifulSoup
- PostgreSQL
- Redis
- TypeScript
- Tailwind CSS
- HeroUI
- Supabase
- Jest
- AWS S3
- CloudFront
- Claude AI API
- AWS Lambdas
Past
FormHealth Website
Collaborated with FormHealth to develop their new design system and rebuild the front-end implementation within their existing Next.js, Prismic, and Storybook stack. Handled ongoing website maintenance typical of a small startup environment while working closely with the marketing team to support their growth initiatives and content needs.
- Next.js
- TypeScript
- Tailwind CSS
- Prismic
- Storybook
SOLIDWORKS Cloud Platform
Collaborated closely with the platform services team to implement comprehensive design system technology and documentation systems across multiple SolidWorks products. Managed product feature requests, system maintenance, and cross-platform integration work spanning the entire SolidWorks ecosystem including Apps for Kids, xDesign, and 3DEXPERIENCE tools. Contributed to prototyping new product concepts while maintaining consistent user experience standards through shared component libraries and design tokens. Handled the critical but often invisible work of keeping multiple products aligned—from updating design system documentation to coordinating feature rollouts across different teams and ensuring seamless integration between desktop and cloud-based applications.
- Angular.js
- JavaScript
- Node.js
- Express.js
- CSS
Apps for Kids Platform
Collaborated on rebuilding the front-end and back-office systems for SolidWorks' educational platform that introduces children ages 4-14 to 3D design and engineering concepts. Redesigned and implemented the user interface to balance child-friendly accessibility with educational depth, while developing comprehensive admin tools for educators to manage classrooms and track student progress. Built new content management features and newsletter tools, creating a seamless learning environment that serves students worldwide while maintaining COPPA compliance for child safety.
- Angular.js
- JavaScript
- Node.js
- Express.js
- Docker
- CSS
- Cassandra
